BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240214T070242Z
LOCATION:Darling Harbour Theatre\, Level 2 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T093000
DTEND;TZID=Australia/Melbourne:20231212T124500
UID:siggraphasia_SIGGRAPH Asia 2023_sess209_papers_661@linklings.com
SUMMARY:VET: Visual Error Tomography for Point Cloud Completion and High-Q
 uality Neural Rendering
DESCRIPTION:Technical Papers\n\nLinus Franke, Darius Rückert, and Laura Fi
 nk (Friedrich-Alexander Universität Erlangen-Nürnberg); Matthias Innmann (
 NavVis GmbH); and Marc Stamminger (Friedrich-Alexander Universität Erlange
 n-Nürnberg)\n\nIn the last few years, deep neural networks opened the door
 s for big advances in novel view synthesis. Many of these approaches are b
 ased on a (coarse) proxy geometry obtained by structure from motion algori
 thms. Small deficiencies in this proxy can be fixed by neural rendering, b
 ut larger holes or missing parts, as they commonly appear for thin structu
 res or for glossy regions however still lead to very distracting artifacts
  and temporal instability. In this paper, we present a novel neural render
 ing based approach to detect and fix such deficiencies. As a proxy, we use
  a point cloud, which allows us to easily remove outlier geometry and to f
 ill in missing geometry without complicated topological operations. Keys t
 o our approach are (i) a differentiable, blending point-based renderer tha
 t can blend out redundant points, as well as (ii) the concept of Visual Er
 ror Tomography (VET), which allows us to lift 2D error maps to identify 3D
 -regions lacking geometry and to spawn novel points accordingly. Furthermo
 re, (iii) by adding points as nested environment maps, our approach allows
  us to generate high-quality renderings of the surroundings in the same pi
 peline. In our results, we show that our approach can significantly improv
 e the quality of a point cloud obtained by structure from motion and thus 
 increase novel view synthesis quality. In contrast to point growing techni
 ques, the approach can also fix large-scale holes and missing thin structu
 res effectively. Rendering quality outperforms state-of-the-art methods an
 d temporal stability is significantly improved, while rendering is possibl
 e with real-time frame rates.\n\nRegistration Category: Full Access, Enhan
 ced Access, Trade Exhibitor, Experience Hall Exhibitor
URL:https://asia.siggraph.org/2023/full-program?id=papers_661&sess=sess209
END:VEVENT
END:VCALENDAR